The Concept of Validity

Validity is a measure of whether the selection method appropriately, meaningfully and usefully:

(a) predicts candidates’ behaviour, such as performance, when they become a job-holder. This kind of validity is called criterion-related validity in that criteria such as performance are used to measure the validity of the selection method, or

(b) represents an important aspect of the job such as knowledge of geography, driving laws or engine mechanics. This is called content validity, or

(c) measures a construct (an underlying human trait or characteristic, such as integrity or honesty) that is believed to be an important aspect of the job. This is called construct validity.

Table 2
Validity

One other consideration with respect to validity is what is called a validity coefficient. This coefficient is expressed as either a positive or negative number where a minus 1.00 means there is a perfect, negative correlation in the validity between the test and what it is trying to predict and a plus 1.00 means there is a perfect, positive correlation. Correlation numbers equal to or greater than plus or minus .35 are considered to represent satisfactory validity. It is important to determine the validity coefficient when considering a selection method.

 While we can offer no “silver bullet” for hiring the best candidates, using selection methods with a high degree of validity and reliability helps remove the tarnish!

Industrial and organizational psychologists have been wrestling with the issue of validity, as it relates to selection, for almost a century (see the following note on the Boston Railway Company). In the course of their research, certain findings have emerged that are relevant to us in developing or selecting “best practice” tools. For the purpose of our consideration of this topic, the most important findings are as follows:

• Where a detailed job analysis has not been performed and, therefore, there is not general agreement on the content of the job, a criterion-related method is preferred.
• Where a detailed job analysis (such as the National Occupational Standards) has been performed, methods based upon content validity are generally recommended.
• Notwithstanding validation studies using any or all of the above three approaches, tests should have high “face validity.”¹²
• A combination of methods (e.g. background review, structured interviews and tests), as opposed to use of a single method can improve outcomes.
• As a general strategy, lower validity approaches should precede the use of higher validity approaches. Therefore, the use of application forms, résumés and reference checks should precede the use of interviews, which, in turn, should precede the use of tests, assuming the tests have been validated for the bus operator position.
• Understanding and being able to use the method (whether it be reference checking, interviewing or interpreting test results) effectively is crucial to it being a valid process for selection.

 A test or selection technique must appear “on its face” to be meaningful and appropriate to managers, union leaders, employees and applicants.

¹⁰ It is interesting to note that in the 1980’s, OC Transpo of Ottawa was using a perceptual ability test as one of a battery of tests to select drivers. In a 1991 study, conducted by industrial psychologists from the University of Guelph, it was determined, through measurement by an in-service riding check, that the test was predictive of job performance but when using supervisory ratings was not predictive of job performance.

¹¹ In the same OC Transpo study, it was determined that the widely used Metro Seattle Video Test for bus drivers predicted that candidates who scored higher on the test subsequently had fewer passenger complaints when they became drivers.

¹² “Face validity” refers to the notion that the test or selection technique has to appear “on its face” to be meaningful, appropriate and useful to managers, union leaders and employees in the company and to applicants.